Imagine a future in which artificial intelligence develops so rapidly that it consumes about 10% of the world’s SSD output, leaving inventors, gamers, and regular users in a state of panic. That is exactly what Nvidia’s new Vera Rubin AI platform, which was revealed at CES 2026, promises: a profound and frightening change in computing. The hidden cost of AI’s dreams becoming a reality is a worldwide storage shortage that affects everyone.
Vera Rubin: The Beast of AI Awakens
This next-generation platform, which bears the name of astronomer Vera Rubin, incorporates Rubin GPUs, Vera CPUs, HBM4 memory, and enormous NAND storage into supercomputers for models with trillions of parameters. For context storage and inference, each NVL72 rack requires 1,152 TB of SSDs, which is sufficient to store millions of hours of video.
It’s exciting: 13-22 TB/s bandwidth, 3nm technology, 288 GB HBM4 per GPU. However, the realisation that this power has a cost—that Nvidia’s need for NAND may reshape our daily access to technology—brings the emotion home.
The Startling Shortage of Supplies
According to Citi analysts, 34.6 million TB of NAND will be needed for 30,000 Rubin units in 2026 (2.8% of the global supply), and that number will skyrocket to 100,000 units in 2027 at 115.2 million TB, or 9.3% of global production. That’s 16 TB per GPU rack, putting AI ahead of consumer SSDs and possibly raising costs by 20–50%.
The heartbreak: Similar to previous semiconductor crises, everyday SSDs for laptops, PS5, and makers may become scarce. However, it’s evidence of AI’s relentless progress, transforming our digital lives from fantasy to need.
Innovations in Technology Fuelling the Fire
Through partners like Kioxia, Nvidia pairs Rubin with LPDDR5X (1.5 TB per CPU) and PCIe 7.0 SSDs that will reach 100 million IOPS by 2027. Rubin Ultra goes one step farther with 384 GB HBM4E and 600 kW racks for robotics and agentic AI. This is evolution rather than greed, allowing for more intelligent beings to dream, create, and solve problems like never before.
Experts caution that memory will become the new battlefield for AI since Nvidia is claiming 77% of AI wafers.
The AI Crossroads for Humanity
Mixed feelings are evoked by Nvidia’s SSD appetite: wonder at advancement and fear of shortage. AI may surpass consumer technology by 2027, driving advancements in storage such as Micron’s 9550 NVMe SSDs. It is a call to adapt for users and an inspiration for innovators. Vera Rubin is more than just hardware; it’s our future, which demands that we strike a balance between equity and ambition.